Predictors of Job Satisfaction and Intent to Leave among Home Health Workers: An Analysis of the National Home Health Aide Survey. Analytical Approach

08/14/2015

An analytic file was prepared by merging data from the NHHAS and NHHCS using agency identifiers. Due to the confidential nature of selected data elements (e.g., CMS provider numbers) and the need to access non-publicly available data, the merge was performed by the NCHS Research Data Center. Following this merge, a de-identified, aide-level file was made available for use in this study. All analyses were conducted with weighted samples, using a with-replacement design approach. The NHHAS represents a weighted population of 160,720 home health workers. Accounting for missing or anomalous data needed to perform these analyses, a total of 119,500 weighted cases, were included in this study. Compared to cases excluded due to missing data, the sample included in the analysis is not statistically different in terms of job satisfaction or intent to leave job; however, they are more likely to work for a non-profit agency, to hold only one job, to work full-time, to have a pension/retirement, and to receive paid time off (sick leave, personal days, and holidays). Table 1 presents descriptive statistics of the 119,500 weighted cases included in the analysis for a subset of variables.

Bivariate analyses, consisting primarily of frequencies, means, and corresponding tests of significance, chi-square and t-tests, respectively, were conducted to understand how home health worker subgroups--defined by level of job satisfaction and intent to leave their job--differed in terms of the characteristics of the aide, agency, work environment, and organizational culture. These results are summarized in Table 2. Data were analyzed with SAS-callable SUDAAN to account for the NHHAS complex survey design. Chi-square tests and t-tests were performed to determine whether differences in subgroups were statistically significant at the p<0.05 level.

Multivariate Analyses of Job Satisfaction: The Multilog procedure in SAS-callable SUDAAN was used to estimate the multinomial logistic regression models. Multinomial logistic regression, which assumes that the dependent variable is nominal, was used to assess the determinants of job satisfaction among home health workers.1 “Extremely satisfied” is compared to “dissatisfied” and “somewhat satisfied” to “dissatisfied” for the analytical sample of 119,500 weighted cases. Coefficients obtained from the multinomial logistic regression are interpreted as the effect of the independent variable on the log-odds of being in one category as opposed to the base category or reference group. Job satisfaction was coded as three levels. For the purpose of this analysis, “dissatisfied” respondents were identified as the reference group.

View full report

Preview
Download

"JobSatPre.pdf" (pdf, 655.64Kb)

Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®